[Skip to Content]
[Skip to Content Landing]
October 2017

The Expanding Spectrum of Zika Virus Infections of the Nervous System

Author Affiliations
  • 1Department of Neurology, University of Colorado School of Medicine, Aurora
  • 2Department of Medicine, University of Colorado School of Medicine, Aurora
  • 3Department of Immunology-Microbiology, University of Colorado School of Medicine, Aurora
  • 4Department of Neurology, Indiana University School of Medicine, Indianapolis
JAMA Neurol. 2017;74(10):1169-1171. doi:10.1001/jamaneurol.2017.1471

Zika virus was first isolated in 1947 from the blood of a captive rhesus macaque on an elevated platform in the tree canopy of the Zika Forest, presumably infected by mosquitos.1 Human infection with the Zika virus was not documented until the early 1950s, and most reports prior to the current outbreaks in the Americas in 2015 emphasized the relatively benign nature of infection. Reports from Micronesia in 2007 and French Polynesia in 2013 suggested that approximately 80% of infections were asymptomatic and that the remaining 20% of infected individuals developed a combination of rash, fever, arthralgia, conjunctivitis, myalgia, and headache (listed in descending order of frequency).2

The outbreaks in the Americas that began in 2015 were notable both for their size and for the recognition that Zika infection could be associated with severe neurologic complications. The first of the recognized neurologic syndromes was congenital microcephaly complicating Zika virus infection during pregnancy. Microcephaly is now recognized as only 1 possible outcome of “congenital Zika syndrome.” Central nervous system effects of congenital Zika syndrome include brain atrophy; intracranial calcifications; cortical, callosal, and cerebellar malformations; neural tube defects; arthrogryposis; retinal and optic nerve abnormalities; and hearing loss, among others.3 The overall risk of congenital neurologic abnormalities in fetuses and liveborn children of mothers enrolled in the Center for Disease Control and Prevention’s Zika Pregnancy Registry with definite laboratory evidence of Zika virus infection during pregnancy was approximately 17-fold higher than the expected baseline rate in uninfected mothers (50 per 1000 live births [95% CI, 35 to 75 per 1000 live births] vs 3 per 1000 live births), and 15% of mothers with evidence of definite infection occurring during the first trimester of pregnancy had infants born with congenital Zika syndromes.4,5 The pathogenesis and mechanism of neural injury in congenital Zika syndromes remain to be elucidated, although results of experimental studies suggest that the virus can directly infect neural progenitor cells and glial cells, inducing apoptotic and excitotoxic cell death, inhibiting cellular proliferation, triggering neuroinflammation, and reducing both the neural progenitor pool and their subsequent neuronal migration.6,7

Guillain-Barré syndrome (GBS) was the first major nonpregnancy-related neurologic syndrome to be linked to Zika infection. During the Zika epidemics in French Polynesia and the Americas, there was a 10- to 40-fold increase in the incidence of GBS cases compared with the pre-Zika baseline and a risk of approximately 1 case of GBS per 4000 individuals infected with the Zika virus. Risk of Zika virus–associated GBS may increase with increasing age. A retrospective study in French Polynesia suggested that the predominant subtype of Zika virus–associated GBS was acute motor axonal neuropathy.8 A subsequent study from Columbia, confirmed by reports from other countries in the Americas, indicated that acute inflammatory demyelinating polyneuropathy, rather than acute motor axonal neuropathy, was the predominant subtype (seen in approximately 78% of cases).9 In both French Polynesia and Columbia, cases of GBS occurred in close temporal proximity to symptomatic Zika infection (6-7 days).8,9 This short latent period raises the possibility that direct infection or parainfectious mechanisms rather than postinfectious immune-mediated processes may be critical to the pathogenesis of Zika virus–associated GBS9 and is consistent with reports that Zika virus–associated GBS can even occur simultaneously with acute infection.10 Antiganglioside and/or antiglycolipid antibodies, a frequent marker of immune-mediated processes in GBS, have been reported in 31% of cases at disease onset and in 48% at 3 months after infection, but without a consistent glycolipid target.8 Factors accounting for reported variations in Zika virus–associated GBS phenotypes remain unclear and could include viral strain, host factors (including the concomitant presence of antibodies against dengue or other flaviviruses) or technological issues such as electrophysiological techniques and criteria used to classify disease phenotypes. It remains to be determined if the prognosis in Zika virus–associated GBS parallels that of non-Zika–associated disease, and whether treatments proven efficacious in non-Zika virus–associated GBS (eg, intravenous immunoglobulin and plasmapheresis) are equally effective in Zika virus–associated GBS, and how the prognoses of Zika virus–associated GBS and non-Zika virus–associated GBS compare.

Recent reports have described isolated cases of a significantly broader spectrum of potential Zika virus–associated neurologic diseases, including cases of meningoencephalitis and transverse myelitis.11,12 It is critically important to understand both the spectrum and frequency of Zika virus–induced neurologic diseases. The article by da Silva and colleagues13 in this issue of JAMA Neurology represents an initial approach to this problem—a simple survey of the number and frequency of Zika virus–associated neurologic illnesses in broad categories including meningoencephalitis, transverse myelitis, and polyneuropathy. Considerably more sophisticated epidemiologic studies will be needed to establish the frequency of specific syndromes and their associated risk factors. In this study, the authors performed a prospective observational study of consecutive patients presenting to a tertiary care academic hospital specializing in neuromuscular and infectious diseases in the state of Rio De Janeiro, Brazil, between December 2015 and May 2016.13 During the reporting period, they found that 93% of the patients with GBS (27 of 29), 71% of the patients with encephalitis (5 of 7), 67% of the patients with transverse myelitis (2 of 3), and the single patient with chronic inflammatory demyelinating polyneuropathy showed evidence of recent Zika virus infection. They estimated that the frequency of admissions for GBS increased nearly 6-fold compared with the pre-Zika baseline, admissions for encephalitis increased 3.5-fold, and admissions for transverse myelitis remained constant. They found a slightly longer interval between GBS and Zika symptom onset (median, 10 days; range, 4-22 days) than prior studies had noted. They also noted a mixed frequency of GBS subtypes, with 18 of 27 patients (67%) having an electrophysiologic pattern consistent with acute inflammatory demyelinating polyneuropathy, 2 of 27 patients (7%) with acute motor axonal neuropathy, 6 of 27 patients (22%) with acute motor and sensory axonal neuropathy, and a single case of Miller-Fisher syndrome. Some patients had unusual features, including 3 patients with acute motor and sensory axonal neuropathy with brisk reflexes in weak limbs, 3 of 6 patients with encephalitis who also had prominent peripheral nerve involvement, and 4 patients with GBS who showed cranial nerve enhancement on MRI scans.

Because of factors including referral bias, a study such as this cannot determine exact frequencies of Zika virus infection among patients with the syndromes examined, but it may serve to help confirm that Zika can be associated with such syndromes. However, before reaching this conclusion, it is important to understand the basis for a diagnosis of Zika virus infection and the potential pitfalls in establishing a diagnosis of acute Zika virus infection in areas with cocirculating arboviruses such as dengue and chikungunya. The acute clinical syndromes of these viruses overlap considerably and may not be distinguishable; as a result, definitive diagnosis depends on nucleic acid amplification using reverse transcriptase–polymerase chain reaction (RT-PCR) and serologic testing. Amplification of Zika virus RNA by RT-PCR from serum or cerebrospinal fluid (CSF) samples is highly specific for acute infection; however, RT-PCR is relatively insensitive because detectable viral nucleic acid is typically present only in these fluids during the first week after symptom onset. In this study, only 4 of 35 patients had a diagnosis of Zika virus infection made by RT-PCR.13 Recent studies suggest that RT-PCR can detect Zika virus nucleic acid in urine for several weeks after acute infection, but urine samples were not evaluated in this study. For 31 of 35 patients, the diagnosis was via serologic testing. Serologic diagnosis of flavivirus infection is complex because infection with 1 virus can result in heterologous cross-reacting antibodies against other closely related viruses, which is particularly problematic in areas such as Brazil where both dengue and Zika cocirculate. The authors report that only 2 of the 32 patients with serologically diagnosed Zika virus infection (both with GBS) also had dengue virus IgM in serum samples, and in both cases, they had Zika but not dengue IgM detected in CSF samples.13 Unfortunately, the prevalence of dengue IgG is not reported because prior flavivirus infection and associated antibodies may play a role in altering Zika virus infection. In this series, a surprisingly high percentage of cases had Zika virus IgM in both serum and CSF samples regardless of disease phenotype. For example, 23 of 26 patients with GBS (88%) had Zika IgM in both serum and CSF samples. This result is intriguing if confirmed in other series because the presence of virus-specific CSF IgM is usually taken as an indication of intrathecal synthesis and considered to be reflective of the presence of viral target antigens within the central nervous system. It does not appear that dengue CSF IgM was looked for in most patients,13 which would have served as an important specificity control. Improved serologic techniques for Zika virus infection are a subject of active development. The confounding issues raised by serodiagnosis of flavivirus infections remain a considerable problem. Recent studies suggest that assays that detect antibody responses specifically directed against the NS1 protein of the Zika virus14 are sensitive and much more virus specific. These NS1-specific antibodies are also less likely to be confounded by heterologous cross-reactivity than assays such as the IgM antibody-capture enzyme-linked immunosorbent assay used in this study, which may detect cross-reacting antibodies against the viral envelope (E) protein. Until these virus-specific assays are validated, specific diagnosis of acute Zika neurologic disease is optimally based on amplification of Zika virus nucleic acid from CSF by RT-PCR1 or detection of Zika virus IgM in CSF in the absence of IgM against other flaviviruses.2 Detection of Zika virus nucleic acid in serum or urine samples, detection of serum Zika IgM in the absence of other flavivirus IgM, and/or Zika IgG seroconversion between acute and convalescent sera in the absence of antibodies against other flaviviruses may all be considered evidence of acute Zika virus infection, but they provide less direct evidence of a causal association with coexisting neurologic disease.

Flavivirus antibody responses are not just an issue in disease diagnosis; they may in fact be critical in disease pathogenesis. For example, antiviral antibodies can facilitate viral entry into target cells via binding to host cell Fcγ receptors, a process known as antibody-dependent enhancement. In the case of dengue, antibody-dependent enhancement appears to be a potential factor in the development of severe dengue virus disease, which often occurs in patients previously infected with 1 of the 4 dengue virus serotypes who are then subsequently infected with a different serotype. Preexisting antibodies against dengue or another flavivirus from prior infection could conceivably also alter the phenotype of Zika virus disease. For example, mice receiving intraperitoneal transfer of convalescent sera from individuals previously infected with dengue and then challenged with Zika virus exhibited protection from Zika with high doses of transferred dengue immune plasma (200 µL) but had significantly higher mortality when they received low doses (2-20 µL) (79% mortality after 20-µL dengue plasma transfer compared with 7% for control plasma). Similar but less dramatic dose-dependent results were seen after the transfer of West Nile virus convalescent plasma.15 The specificity of the antibodies may be important as well because dengue virus monoclonal antibodies with differing epitope specificities cloned from plasmablasts of patients with dengue virus differ in their capacity to protect mice against challenge with Zika virus.16 Understanding the complexities of flavivirus antibody responses will likely be critical for both diagnosing Zika virus infections and understanding their pathogenesis.

Back to top
Article Information

Corresponding Author: Kenneth L. Tyler, MD, Department of Neurology, University of Colorado School of Medicine, 12700 E 19th Ave, Campus Mailstop B-182, Research Complex II, Aurora, CO 80045 (ken.tyler@ucdenver.edu).

Published Online: August 14, 2017. doi:10.1001/jamaneurol.2017.1471

Conflict of Interest Disclosures: None reported.

Dick  GW, Kitchen  SF, Haddow  AJ.  Zika virus, I: isolations and serological specificity.  Trans R Soc Trop Med Hyg. 1952;46(5):509-520.PubMedGoogle ScholarCrossref
Duffy  MR, Chen  TH, Hancock  WT,  et al.  Zika virus outbreak on Yap Island, Federated States of Micronesia.  N Engl J Med. 2009;360(24):2536-2543.PubMedGoogle ScholarCrossref
Melo  AS, Aguiar  RS, Amorim  MM,  et al.  Congenital Zika virus infection: beyond neonatal microcephaly.  JAMA Neurol. 2016;73(12):1407-1416.PubMedGoogle ScholarCrossref
Cragan  JD, Mai  CT, Petersen  EE,  et al.  Baseline prevalence of birth defects associated with congenital Zika virus infection—Massachusetts, North Carolina, and Atlanta, Georgia, 2013-2014.  MMWR Morb Mortal Wkly Rep. 2017;66(8):219-222.PubMedGoogle ScholarCrossref
Reynolds  MR, Jones  AM, Petersen  EE,  et al; U.S. Zika Pregnancy Registry Collaboration.  Vital signs: update on Zika virus–associated birth defects and evaluation of all US infants with congenital Zika virus exposure—US Zika Pregnancy Registry, 2016.  MMWR Morb Mortal Wkly Rep. 2017;66(13):366-373.PubMedGoogle ScholarCrossref
Tang  H, Hammack  C, Ogden  SC,  et al.  Zika virus infects human cortical neural progenitors and attenuates their growth.  Cell Stem Cell. 2016;18(5):587-590.PubMedGoogle ScholarCrossref
Li  H, Saucedo-Cuevas  L, Shresta  S, Gleeson  JG.  The neurobiology of Zika virus.  Neuron. 2016;92(5):949-958.PubMedGoogle ScholarCrossref
Cao-Lormeau  V-M, Blake  A, Mons  S,  et al.  Guillain-Barré Syndrome outbreak associated with Zika virus infection in French Polynesia: a case-control study.  Lancet. 2016;387(10027):1531-1539.PubMedGoogle ScholarCrossref
Parra  B, Lizarazo  J, Jiménez-Arango  JA,  et al.  Guillain-Barré syndrome associated with Zika virus infection in Colombia.  N Engl J Med. 2016;375(16):1513-1523.PubMedGoogle ScholarCrossref
Siu  R, Bukhari  W, Todd  A, Gunn  W, Huang  QS, Timmings  P.  Acute Zika infection with concurrent onset of Guillain-Barré Syndrome.  Neurology. 2016;87(15):1623-1624.PubMedGoogle ScholarCrossref
Carteaux  G, Maquart  M, Bedet  A,  et al.  Zika virus associated with meningoencephalitis.  N Engl J Med. 2016;374(16):1595-1596.PubMedGoogle ScholarCrossref
Mécharles  S, Herrmann  C, Poullain  P,  et al.  Acute myelitis due to Zika virus infection.  Lancet. 2016;387(10026):1481.PubMedGoogle ScholarCrossref
da Silva  IRF, Frontera  JA, Bispo de Fillipis  AM, Nascimento  OJMd; RIO-GBS-ZIKV Research Group.  Neurologic complications associated with the Zika virus in Brazilian adults  [published online August 14, 2017].  JAMA Neurol. doi:10.1001/jamaneurol.2017.1703Google Scholar
Steinhagen  K, Probst  C, Radzimski  C,  et al.  Serodiagnosis of Zika virus (ZIKV) infections by a novel NS1-based ELISA devoid of cross-reactivity with dengue virus antibodies: a multicohort study of assay performance, 2015 to 2016  [published online December 15, 2016].  Euro Surveill. 2016;21(50):30426. doi:10.2807/1560-7917.ES.2016.21.50.30426PubMedGoogle ScholarCrossref
Bardina  SV, Bunduc  P, Tripathi  S,  et al.  Enhancement of Zika virus pathogenesis by preexisting antiflavivirus immunity.  Science. 2017;356(6334):175-180.PubMedGoogle ScholarCrossref
Kam  YW, Lee  CY, Teo  TH,  et al.  Cross-reactive dengue human monoclonal antibody prevents severe pathologies and death from Zika virus infections.  JCI Insight. 2017;2(8):92428. PubMedGoogle ScholarCrossref